sign in sign up
<<
Home , Adityawarman, Adityawarman Museum, Banjarmasin, Banten, Batang Hari River, Bengkulu (city)

Map I. Southern , showing location of six of the early inscriptions EARLY POLITIES IN SOUTHERN SUMATRA: SOME PRELIMINARY OBSERVATIONS BASED ON ARCHAEOLOGICAL EVIDENCE*

E. Edwards McKinnon

Introduction

While it is known from historical sources that certain polities existed in southern Sumatra during the late first and early second millennium AD, it has been difficult to establish conclusive archaeological evidence concerning them. These circumstances have arisen for several reasons, among the most important being lack of any systematic search for archaeological remains in the region, the terrain itself, and insufficient knowledge of ancient demography. Most archaeological sites in the area have been known since the nineteenth or early twentieth centuries, when they were reported by travelers or colonial administrators, such as John Anderson or L. C. Westenenk. Since then, however, much of the earlier evidence has been destroyed. Sites have been plundered for building materials and have all but disappeared, and images have been removed to museums for safekeeping without accurate records being made of their original locations. There are a limited number of key sites in the Musi and valleys which have not been examined by professional archaeologists for over half a century and which may still reveal valuable information about earlier times. In addition, habitation sites have been recognized more recently in the lower reaches of both rivers which will enable archaeologists to make a much more comprehensive assessment of cultural interaction in the region than has been possible hitherto. The first professionally conducted excavation was carried out in only in 1974. Since 1982, however, there have been limited but extremely useful excavations at Muara . This article will attempt to bring together the archaeological evidence that has been found thus far in the Musi and Batang Hari river valleys and suggest what light these shed upon the nature of the polities existing in the region from the seventh to the fourteenth centuries.

The Late First Millennium

Evidence from the Inscriptions

Six inscriptions, t dating to the latter part of the seventh century AD are the earliest epigraphical evidence of sophisticated social and political *1

* A first draft of this article was presented in May 1984 at a symposium on Southeast in the ninth and fourteenth centuries at the Australian National University, Canberra. 1. Now seven, a further "curse" inscription having been found very recently

1 2

organization in southern Sumatra. They indicate the existence here of a major center of power in the region at that time, that of the ancient Malay thalassocracy of Srivijaya. Of these inscriptions, three are to be found in the vicinity of Palembang, the modern provincial capital of Province on the lower course of the . One came to light at Talang Tuo, some eight kilometers west of Bukit Seguntang across the swamp known as Putih Kuku, and a second at Kedukan Bukit,2 just to the south of Bukit Seguntang. The third, bearing the text of a curse, was found at Sabokingking to the west of the modern , on an area of higher ground lying between Gede ing suro and the Sei Buah, immediately to the south of a tank or pond known as Telaga Batu. There are three other major curse inscriptions relating to Srivijaya, all bearing basically similar texts, but that of Sabokingking is the most sophisticated.3 Indeed, both the text of the curse and the form of the great naga stone of Sabokingking are the most elaborate presentation of any of the Srivijayan inscriptions. One cannot do better than repeat de Casparis* translation of Stutterheim’s description of this stone: The stone is hooded by seven cobra heads with flat, round jewels, broad necks and neck wrinkles, which, at the back of the stone gradually merge into the flat surface of the stone. On the front side, 28 lines of script are visible, badly weathered and illegible but for a few aksaras. The inscribed part of the stone is separated from the roughly flattened underpart by a protruding horizontal ledge, which descends a little before joining the surface of the stone; thus a groove is formed, passing in the middle part into a spout in the form of a Yoni.4 The inscription comprises two parts, a short introductory text in a dialect that is imperfectly understood and which has so far defied clear translation, followed by a second, longer text in Old Malay.5 The lettering is a form of Tamil Grantha script, the earliest form of writing known in .

at Desa on the Way Sekampung in South . Information from Drs. Bambang Budi Utomo. 2. Westenenk refers to "door hen gegraven kanaal Kedoekan Boekit," L. C. Westenenk, "Bukit Segoentang en Goenoeng Mahameroe uit de Sedjarah Melayoe," Tijdschrift voor Indische Taal-, Land- en Volkenkunde (hereafter TBG) 63 (1923): 212-26, 216. Keduk is from Javanese, mengeduk, to dig or scoop out. 3. J. G. de Casparis, "The Old Malay Inscription of Telaga Batu (South Sumatra)," in his Prasasti II (: Masa Baru, 1956), pp. 15-46. See also, Boechari, "An Old Malay Inscription of Srivijaya at Palas Pasemah (South Lampung)," in Pra Seminar Penelitian Srivijaya, ed. Drs. (: Pusat Penelitian Purbakala dan Peninggalan Nasional (hereafter Puspan], 1979), pp. 18-40. 4. De Casparis, "Old Malay Inscription," p. 15, translated from Jaarboek van het Koningklijke Bataviasche Genootschap III [1936], p. 198. The existence of this particular stone raises questions regarding the nature of ophiolatry in Sumatra during the seventh century. It has, however, a long association with in . The Buddhism of Sumatra was initially , which culminated in Vajrajana or left-handed Tantrism of the Kalacakra, latterly merging with Saivism. 5. Boechari, "Old Malay Inscription," p. 23. 3

Plate 1. The Great Naga Stone of Sabokingking

(Courtesy: Muaeun Nasionel, Jakarta} Map II. Musi River Basin 5

The imprecation formula comprises three parts: an invocation to the Gods, a curse to evildoers, and finally a blessing (tantramala) to those who remain loyal to the (Chief). The introductory portion of the inscription appears to pronounce a shorter version of the fuller, Old Malay, text.8 The Saboking- king version is unique, in that here the Datuk refers to himself as "Aku,n a reference which is thought to imply the physical presence of the Chief him­ self. 6 7 8910 Both the text and the physical form of the stone indicate that the vassals of the Datuk were required to drink the curse, or, more precisely, to drink sanctified water which had been poured over the stone and which then embodied the spirit of the imprecation. Both de Casparis8 and Kenneth Hall8 suggest similarities with such practices known from India and elsewhere in Sumatra. Later, in the eighteenth century William Marsden, writing about the laws and customs of the Rejang, who live around the headwaters of the Musi, noted an interesting form of oath taking or makan sumpah (to swallow an oath): The place of greatest solemnity for administering an oath, is the krammat or burying ground of their ancestors, and several superstitious ceremonies are observed on the occasion. The people near the sea-coast, in general, by long intercourse with the , have an idea of the Koran, and usually employ this in swearing, which the priests do not fail to make them pay for; but the inland people keep, laid up in their houses, certain old reliques, called in the Rejang pesakko, and in Malayan, sactian, which they produce when an oath is to be taken. . . . They consist of an old rusty , a broken gun barrel, or any ancient trumpery, to which chance or caprice has annexed an idea of extraordinary virtue. These they generally dip in water, which the person who swears drinks off, after having pronounced the form of words before mentioned. The pangeran of Sungai-lamo has by him certain bullets, which had been steeped in water, drunk by the Sungei etam chiefs, when they bound themselves never to molest his dis­ tricts: which they have only done since, as often as they could venture it with safety, from the relaxation of our government. . . . In all these ceremonies, they burn on the spot a little gum ben­ zoin. The above quotation is interesting in that it illustrates the tenacity of such beliefs in what were apparently similar circumstances, binding insubordinate chiefs to an acceptable code of conduct. In the ultimate instance, however,

6. Ibid., p. 25. 7. J. G. de Casparis, in discussing "Some Notes on the Epigraphical Heritage of Srivijaya," SPAFA Digest 3, 2 (1982): 29-34. 8. De Casparis, "Old Malay Inscription." 9. See K. R. Hall, "State and Statecraft in Early Srivijaya," in Explorations in Early Southeast Asian History: The Origins of Southeast Asian Statecraft, ed. Kenneth R. Hall and John K. Whitmore (Ann Arbor: University of Michigan Center for South and Southeast Asian Studies, 1976), pp. 89-91. 10. William Marsden, The History of Sumatra (: Oxford in Asia Historical Reprints, 1966), pp. 242-43. History, it seems, tends to repeat itself. 6

the final decision would no doubt depend upon the ruler’s superiority in arms and his ability to enforce his will on his vassals.1'' Two of the other curse inscriptions were found at Kota Kapur on the island of Bangka and at Karang Berahi, above Pamenang on the Batang Merangin, a tributary of the Batang Hari in upper Jambi (Nos. 2 and 3 on Map I). Latterly, at Palas Pasemah,11 12 another version of the curse has been found on a ridge overlooking the Way Pisang, a tributary of the Way Sekampung, the largest river in southern Lampung (No. 4 on Map I). The Way Pisang rises on the eastern slopes of Gunung Rajabasa, a peak which commands the Selat Sunda, the straits lying between Lampung, Sumatra, and in West . The various sites of the curse inscription would appear to have been focal points in their respective localities, perhaps already the seats of local chiefs rather than boundary markers of an expanding polity, where people from surrounding areas could be assembled to swear allegiance to their overlords. These locations take on particular significance in an understanding of both the political and economic infrastructure of South Sumatra in the late seventh century. Two, in fact, are of particular importance. Kota Kapur on the island of Bangka is strategically situated to command the major shipping route from Java and the spice islands of Eastern Indonesia to , India, and the West. A settlement at Kota Kapur, together with Sungsang on the Musi estuary (see Map II) could have established a complete stranglehold on all East-West shipping. Karang Berahi, situated on the threshold of a major alluvial gold-bearing region in upper Jambi was also of strategic importance, commanding as it does, access to a large part of the gold-producing areas of the Batang Merangin and Kerinci. Although upper Jambi may have been economically important, however, it could only realize its full potential through an outlet to a market on the major trading route along the east coast of Sumatra. Thus, whoever commanded the lower Batang Hari would have also, no doubt, been able to influence, either directly or indirectly, what went on upstream. It is here that Chinese records may be able to shed some light on the subject. The Hsin Tang Shu (New Tang History) mentions a country referred to as Kompe which sent a mission to China in the year 662 AD, some twenty years before the Srivijayan expansion into Jambi. It may be no coincidence that there is a Kompeh on the lower Batang Hari, which may have been the port of Melayu.13 In order to reach Karang Berahi by water, a Srivijayan expedition would have had to have passed Kompeh and (Muara) Jambi, probably subduing

11. De Casparis notes that it is possible that the imprecations were not considered sufficiently convincing to all subjects of the Datuk and, for that reason, real punishment is also referred to. See de Casparis, "Old Malay Inscription,” p. 31. 12. Palas Pasemah, Kecamatan Palas, Kabupaten Lampung Selatan, is said to have been known formerly as Pelabuhan Ratu or Tanjung Merindu. The name "Pasemah” was apparently adopted as late as the 1930s with the arrival of immigrants from the Pasemah area of South Sumatra. The site would appear, however, to have been inhabited in earlier times, as numerous blue and white (Ming period) sherds are to be found in this area. 13. E. Edwards McKinnon, "A Brief Note on Muara Kumpeh Hilir: An Early Port Site on the Batang Hari?" SPAFA Digest 3, 2 (1982): 37-40. Plate 2. The Kota Kapur Stele

(Courtesy: Museum Nasionel, Jakarta)

Plate 3. The Palas Pasemah Inscription 8

Plate 5. The Karang Berahi Inscription (detail)

One of the mortises cut into the shoulder of the stone. 9 its chief in the process. By overcoming the forces of Kompeh/Jambi, the Datuk would no doubt have acquired a certain amount of "mystical prowess" in addition to any acknowledgement of physical military superiority. The relative significance of Karang Berahi in the Srivijayan is reflected in the form of the inscription stone itself. Second in importance only to that of Sabokingking, the Karang Berahi stone is carved in the shape of a short bullet with one flattened side bearing the lines of the inscription. The stone itself is relatively plain. It has, however, one peculiar feature. On the right and left shoulders at the top of the stone are two mortices, cut, it would seem, to receive the tenons from some missing part (see Plate 5 ) conceivably a frieze of serpents’ heads formed from either stone or wood in the same concept as that of the Sabokingking inscription. The Karang Berahi inscription is the earliest datable archaeological document in the Batang Hari basin. The texts of the inscription at Karang Berahi, Kota Kapur, and Palas Pasemah are very similar. The Kota Kapur stele, which bears the date Saka 608 (AD 686), is distinguished by reference to it being set up at the moment when the army of Srivijaya was about to commence an expedition to the land of Java, which had not submitted to Srivijaya. 14 The texts also bear Tantrik overtones. Of the other two Srivijayan inscriptions, that at Kedukan Bukit, which bears a date equivalent to 682, records details of a naval and military expedition at a place named as Minanga Tamwan15 by the Datuk of Srivijaya. By means of this expedition he obtained Siddhyatra or mystical prowess. The toponym "Minanga Tamwan" in this inscription might well refer to Pamenang (the name means victor or victorious), just downstream from Karang Berahi.14 1615 17 The modern name Pamenang is possibly a derivation or variant of minanga or menanga. The word tamwan may conceivably refer to the "mines" or tambang of the area. The 682 inscription may be read as "Dapunta Hyang departed from Minanga tamwan [dapunta hiyang marlapas dari minanga tamwan] while taking with him a [naval] force of 20,000 men and 1,312 infantry." If he arrived from Palembang at this point with separate forces, he could have executed a classical pincer movement, moving upstream along the Batang Hari after a relatively long sea and river journey, with the infantry moving overland across the watershed between the Air Rawas, a major tributary of the Air Musi, and the Batang Asai/Batang Tembesi basins, which provides the modern link between Surulangan in South Sumatra and Sarolangun and Bangko in Jambi province. 17

14. Boechari gives comparative Romanized transcriptions of Kota Kapur, Karang Berahi, and Palas Pasemah in Old Malay. Boechari, "Old Malay Inscription," pp. 38-40. 15. There have been various suggestions regarding the location of Minanga Tamwan including one as far away as . See, for example, G. Coedes, "A Possible Interpretation of the Inscription at Kedukan Bukit (Palembang)," in Malayan and Indonesian Studies, ed. John Bastin and R. Roolvink (Oxford: Clarendon Press, 1964), pp. 24-32. 16. But see Boechari, who suggests that the area subjugated in 682, the name of which begins with a ma, might be upper Jambi. In brief, I am suggesting a connection between the Kedukan Bukit inscription of 682 and that of Karang Berahi of 686. 17. This possibility came to mind with the discovery in 1980 of an important archaeological site on the headwaters of the Sungai Tingkip, a north bank 10

If my assessment of the situation in 682 is correct, the upriver areas both on the Musi and the Batang Hari had by then already assumed considerable economic and political significance. Although it is as yet impossible to ascertain the date at which gold exploitation in these areas commenced, it seems reasonable to suggest that the Srivijayan expansion of the late seventh century was to seize known resources commanded by various local chiefs in their respective areas, and so establish a central authority. How permanently successful this move was at the time is not clear, for there are no more known written testaments of this period. The remaining inscription, that at Talang Tuo, reveals the establishment of a park or garden of fruit trees named Sriksetra by a chief named Jayanaga or Jayanasa. "is It bears the date Saka 606 (AD 684). Recalling the form of the Sabokingking inscription, a broad serpent’s hood surmounted by seven naga heads, is it perhaps no coincidence that the name of the Datuk given at Talang Tuo could be interpreted as Jayanaga, the "victorious serpent"? 19 We have no idea how long the authority of Jayanaga (or Jayanasa) actually lasted. All six inscriptions cover only a very short span of five years (682-686), after which there is no other epigraphical evidence relating to Srivijaya available from southern Sumatra. But although other datable archaeological evidence has yet to be discovered, the "curse" would appear to have been effec­ tive. With his back secured, by the end of the century Jayanaga, or his successor as Datuk, had expanded his mandala to incorporate on the west coast of the . The curse inscriptions indicate that centrifugal forces were an ever-present threat to the stability of the Srivijayan state. These were, no doubt, especially acute in the relative isolation of upper Jambi with its high economic potential. Indeed, such areas were probably virtually independent and were required only to render tribute on a regular basis. In all probability, the degree of absolute control of an area relaxed by direct proportion to the distances from the center of power. It is unlikely that any relatively sophisticated system of administration had evolved. 1819

tributary of the Air Rawas, between Bingin Teluk and Surulangan. Such watersheds were by no means impassable. The upland regions of Sumatra have been honeycombed by trails linking one valley or river basin with another for millennia. 18. Nilakanta Sastri, quoting Coedes in translation, gives: "the garden (named) Sriksetra was made under the direction of His Majesty Sri Jayanasa [Jayanaga?]." K. A. Nilakanta Sastri, History of Sri Vijaya (Madras: University of Madras, 1949), p. 114. More recently, however, Boechari has suggested that the garden was founded by a of Dapunta Hiyang, Sri Jayanasa. Boechari, "Old Malay Inscription," pp. 26-27. I have assumed, for reasons given below, that the name may, in fact, have been Jayanaga. 19. There is, after all, the Indianized marga or clan name Sinaga, found among the Pardembanan and Simalungun peoples of northeast Sumatra and "Naga" does appear as a component part of personal names in Southern India. The form of the great naga stone of Sabokingking may thus reflect not only Buddhistic beliefs, but might conceivably also represent the datuk as protector and overlord of his realm. In Buddhism naga are regarded as water spirits. The whole aquatic complex at Karanganyar (see below) would be entirely fitting to such a concept. For an alternative opinion, see Coedes, who takes the name to be Jayanas'a, "destroyer of Jaya": Coedes, "Possible Interpretation." 11

Plate 6. The Bukit Seguntang Buddha (detail)

Now proven to date from 6th-7th century9 this is therefore roughly contemporary with the inscriptions of Sr1v1jayas 12

Plate 8. Four-armed Avalokitesvara from Bingin Jungun (Ulu Musi)

Similar in style to a 5th century image from Sri Langkef it is probably contemporary with the Bukit Seguntang Buddha. 13

There is one other datable item of evidence for early social and religious development in the Jambi area, in addition to the Karang Berahi inscription. This is the standing Buddha image recovered from Solok (Sipin) near Jambi, which bears an inscription "dan acaryya syuta," datable on epigraphical grounds to the eighth century AD. 20 The religious remains from this period in both the Musi and the Batang Hari basins (although limited to a few images and scattered bricks, very likely foundations of brick and timber shrines) are sufficient to give an indication of the spread of Buddhism throughout the area. The lack of large brick- or stone-built shrines at this time is perhaps not surprising. In India, stone temples began to evolve only in the sixth and seventh centuries. Although in Sumatra, unlike Java, stone-built monuments are virtually unknown, it would appear that stone from earlier structures was reused in later building at a sixteenth century Islamic grave Gede ing suro in east Palembang, where a stone Boddhisatva image was discovered in 1936.^1 a similar occurrence may also have taken place at Candi Angsoka, where several pieces of ashlar masonry are to be seen imbedded amongst the brick rubble. Mahayana Buddhism would appear to have diffused into both the Musi and Batang Hari hinterlands at a fairly early date. Evidence of this upstream movement of religious influence appears at a number of sites on the Musi. For example, at Bingin Jungun, on the middle Musi between Muara Lakitan20 2221 and Muara Kelingi, 2324 a four-armed Lokesvara image influenced in style by fifth century Sinhalese sculpture was discovered in the early years of this century.2* Bingin Jungun, which occupies a strategic point at a bend in the river, is said formerly to have been the head of navigation for larger vessels on the Musi, due to the presence of a cataract. A second seated, and presumably

20. Boechari, "Old Malay Inscription," p. 26. 21. On the discovery of the image, see F. M. Schnitger, The Archaeology of Hindoo Sumatra (Leiden: Brill, 1937), p. 2. I am grateful to Dr. Jacques Dumarcay for the suggestion that the dressed stonework may have come from an earlier shrine. I agree. Several of the "Islamic" sites in Palembang, such as those at Telaga Batu, Gede ing Suro, Candi Angsoka, and the graves on Bukit Seguntang would appear to legitimize themselves by their establishment at or on earlier Hindu Buddhist sacred sites, a common enough occurrence throughout the Islamic world and in . 22. Muara Lakitan, some 300 km above Palembang, stands at the confluence of the Air Musi and the Air Lakitan. A trail from the headwaters of this river gives access to the Lebong (gold-producing) area via Tapus and the Ketahun river valley ( province). 23. Muara Kelingi stands about 330 km above Palembang, at the confluence with the Air Kelingi which rises on Gunung Kaba, in the Rejang Lebong area (Bengkulu province). It was the head of navigation for larger river boats during the colonial period. Rotan and damar were important products in the nineteenth century. 24. Despite the apparent fifth century Sinhalese stylistic affinities, it is likely that this image is a contemporary of the Bukit Seguntang Buddha, once thought to date from the third to fifth centuries, but now satisfactorily proven to date from the seventh century. See Nik Hassan Shuhaimy, "The Bukit Seguntang Buddha, a Reconsideration of Its Date," Journal of the Malayan Branch of the Royal Asiatic Society (hereafter JMBRAS) 52, 2 (1979): 33-40. 14

Map III. Palembang. Antiquities: Site Map

1. Talang Kikim (Edwards McKinnon 1979) 2. Bukit Seguntang (Westenenk 1923, MacLaine Pont 1929) 3. Kampung Bukit (Perquin 1928) 4. Lebak Keranji (Bronson & Wisseman 1974, 1976) 5. Kedukan Bukit (Westenenk 1923, Wellan 1937) 6. Candi Walang (Westenenk 1923, Wellan 1937) 7. Candi Angsoka (Westenenk 1923, Wellan 1937) 8. Ganesa image (Suara Karya 1983) 9. Air Bersih (Bronson & Wisseman 1974, 1976) 10. Geding Suro (Schnitger 1936, 1937, Wellan 1937) 11. Sabokingking (Wellan 1937) 12. Telaga Batu (Wellan 1937) 13. Gunung Mahameru (Westenenk 1923) 14. Earthwork ? (MacLaine Pont 1929) 15. Talang Tuo (Westenenk 1923) 16. Karanganyar (Manguin 1984) Not included in the above are the various findspots of bronze images such as those recovered near Batu Hampar in 3 Ilir and in the Komering, near the confluence with the Air Bandung Besar. 15 later, but unfinished, Buddha image was also recovered from Tebing Candi on the opposite bank of the river.

Karanganyar

The concentration of major inscriptions around the city of Palembang, on the lower Musics has long suggested that this area formed the center of the Srivijayan thalassocracy. When excavation in the mid-1970s failed, however, to reveal evidence of habitation in the Palembang area earlier than the thirteenth century, doubts were expressed regarding its association with Srivijaya, which, as is also attested by foreign historical records, flourished in the region of southern Sumatra between the seventh and twelfth centuries.25 282627 2930 After the first identifiable traces of late first millennium refuse came to light in 1978 in the form of sherds of Chinese greenware at Bukit Seguntang and at Talang Kikim, 27 interest again centered to the west of the city of Palembang, the area where several archaeological remains had also been found during the late 1920s. Then in July 1984 what appears to be an elaborate and extensive complex of canals, tanks, and habitation sites was discovered at Karanganyar, on the north bank of the Musi.88 This discovery seems to substantiate a folk memory recorded by Westenenk as long ago as 1923, when he noted that the local people still quoted in their own dialect, "Dari Pelembang lame sampai Selikoe Batang Hari Leke" (from Palembang Lama to Selikoe Batang Hari Leko) to indicate the extent of a formerly densely populated centrum. 29 The Karanganyar site in 36 Ilir, west of the modern city of Palembang, lies to the south of Bukit Seguntang and Lebak Keranji, an area of firm ground which has yielded traces of fifteenth to sixteenth century occupation and the only known site of ancient glass bead manufacture in Sumatra.38 It appears to be bounded on the north by the line of a canal known as the Sei. Suak Bujang which runs roughly east-west, linking the Sei. Kedukan Bukit above its con­ fluence with the Sei. Tatang in 35 Ilir with the Sei. Rambutan, a stream which

25. For a recent analysis of environmental and spatial relationships in this region, see John N. Miksic, "Penganalisaan Wilaya dan Pertumbuhan Kebudayaan Tinggi di Sumatera Selatan," Berkala Arkeologi 5, 1 (1984): 9-24. 26. The surveys and excavations carried out in 1974 are discussed in: T. Aswar and B. Bronson, Report on the 1974 Sumatra Expedition (Jakarta: Puspan, 1975); B. Bronson, "A Lost Kingdom Mislaid, A Short Report on the Search for Srivijaya," Bulletin of the Field Museum of Natural History [Chicago] 46, 4 (1975): 16-20; B. Bronson and J. Wisseman, Archaeological Research in Sumatra: 1974: A Prelim­ inary Report (Jakarta: Puspan, 1974), mimeograph; B. Bronson and J. Wisseman, "Palembang as Srivijaya. The Lateness of Early in Southern Southeast Asia," Asian Perspectives 19, 2 (1979): 220-39. 27. E. Edwards McKinnon, "A Note on the Discovery of Spur-Marked Yueh-type Sherds at Bukit Seguntang, Palembang," JMBRAS 52, 2 (1979): 41-46. 28. Kompas, March 5, 1985. 29. Westenenk, "Boekit Seguntang," p. 223. 30. Bronson and Wisseman, "Archaeological Research" reported finds of glass slag and beads, together with Chinese blue and white sherds of the Ming period (C14-17) at Lebak Keranji. 16 enters the Musi some distance to the west, between Karanganyar and the Sei. Lamidaro. It is the edge of the higher ground to the south of the Sei. Suak Bujang, barely one kilometer from the north bank of the Musi, which has recently yielded evidence of an imposing complex of carefully laid out tanks and canals concentrated around a rectangular "island” some 310 x 230 m or some 7 hectares in extent. 31 At the present time both tanks and canals are silted up. These tanks and parts of the canals are now largely converted to sawah by the modern inhabitants of the area. The road from 35 Ilir to Gandus, which runs on a bund roughly parallel to the Sei. Suak Bujang, cuts across the southern part of the complex, providing easy access to it. Unfortunately, due to this facility, modern development is rapidly encroaching upon the site which, until a few years ago, was open fields and rubber gardens. Despite the encroachment of modern building, there is relatively little soil disturbance. Seasonal cultivation is limited mainly to the sawah, although the higher ground between the tanks and the Sei. Suak Bujang, and from there to a new road known as Jalan Pam, a few hundred meters to the north, is now being cultivated for cassava and other vegetable production. This area was formerly under neglected rubber gardens and belukar which made access to the area on extremely difficult, but now most of the old rubber has been cleared and succeeded by lalang or cultivation. There is little on the surface to indicate traces of former habitation. 32 However, during a visit to Karanganyar in August 1984 spoil from two recently dug wells at the northern end of the "island" revealed traces of a habitation layer some one to two meters below the present surface. Sherds of late first millennium Chinese stoneware, including fragments of green-glazed "dusun" jars and "patch" bowls, poorly fired locally produced earthenware, bone fragments, and charcoal have been brought to the surface. The charcoal should prove useful in providing an accurate assessment of the dating of the site by C14 analysis. An important factor which would appear to have a significant bearing, both upon the dating and nature of the Karanganyar site, is the original location of the Kedukan Bukit inscription (see above p. 9) which was discovered some sixty years ago only a short distance away to the northeast, on the banks of the Sei. Kedukan Bukit. The exact initial site of this stone is no longer known. If, however, this inscription had stood at or near the entrance to the Sei. Suak Bujang, the nature of text would suggest that it might have been set up to mark the access to a royal complex. From this, it could be inferred that the Karanganyar site has indeed a direct connection with the late seventh century Kadatukan of Srivijaya. There are traces of buried brickwork comprising broad wafer-like bricks at a depth of about 30 cm near the center of the main "island." Unlike the brick remains on Bukit Seguntang, these may be largely undisturbed and are apparently part of walls or foundations of a building or buildings. A large, clear, rock crystal bead, approximately 20 mm in diameter has also come to light in the same area. The central hole of the bead has been bored into the crystal from either side, meeting at a slight angle in the middle, in the fashion of many ancient stone beads from India. Although ancient 3231

31. Information from Dr. P-Y. Manguin. 32. This accords with the impressions of Professor Kenneth Hall who, as a graduate student, assisted Dr. Ben Bronson on a survey of the area in 1974. See Hall, "State and Statecraft," pp. 61-105. 17 beads of agate and carnelian are not uncommon in Sumatra, a rock crystal bead of such dimensions is relatively rare. To the north and south of the central "island" feature are rectangular tanks, linked by canals along the east and west sides. In the center of the northern tank is a small island about 40 m square, the surface of which is littered with fragments of ancient wafer-like bricks. To the east of the main island, and separated from it by a broad shoulder of dry land, now occupied by a farmyard-like group of brick buildings, is an expansive area of sawah some 750 m square and linked to it by a narrow canal, again with a small island in the- center. Some distance to the southwest of the main feature, and again linked to it by a canal which runs from the bottom western point of the most southerly tank, is yet another smaller tank with a regular island in the center some 30 m square, known as Pulau Nangka. A short distance to the west of the main feature, a branch of the Sei. Suak Bujang appears to link the central "island" complex with the canal. The actual dimensions and details of the whole complex will only be revealed when it is remapped from aerial photographs and features are carefully checked out on the ground. The whole area may, however, exceed 100 hectares in extent. Due to the low-lying nature of the ground, the higher parts are probably little more than one or two meters above the surface of the water in the surrounding sawah. Enough remains to be seen on and in the ground to suggest that the location of an ancient seat of power in thePalembang area soon will no longer be a mystery. It is possible that some large-scale excavation will be required to give a clear picture of the site, but the rewards should well compensate the amount of effort required. The presence, in quantity, of Tang period Chinese ceramics should also dispel any remaining doubts that the late first millennium pieces collected before the Japanese occupation from the Palembang area by Van Orsoy de Flines for the museum in Batavia could have come from South Sumatra. The inscriptions and the habitation sites at Bukit Seguntang, Talang Kikim, Talang Kapas, and Karanganyar west of Palembang, together with the Sabokingking inscription from east Palembang begin to give some indication of local spatial organization in the late first millennium. The site of the Pusri fertilizer plant on the north bank of the Musi, downstream of 1 Ilir (Satu Ilir) and the Sei. Buah is said to have yielded finds of late first millennium sherds. 33 It would appear, therefore, that in general, it was the edges of swamps, creeks, and the higher ground (talang) which were inhabited, with perhaps small shrines and situated on the low hills. 34 Remains of brick-built stupa were still to be seen on Bukit Seguntang in the 1920s, and bricks robbed from the remains of ancient monuments 35 formed the foundations of the motor road con­ structed between the area of the Sei. Tatang and Bukit Seguntang. 353433

33. This information comes from Mr. Mundarjito of Universitas Indonesia. 34. The early (C7-8) Avalokitesvara image discovered at the house known as Sarangwati comes from such a site; from the brow of a low hill (see Plate 7). The so-far-undated Candi Angsoka, with its traces of dressed masonry and two massive blocks of limestone, occupies a similar position. 35. P. J. Perquin, "Oudheidkundige onderzoek te Palembang," Oudheidkundig YersLag (hereafter OV) (1928) Bijlage J: 123-28. 18

Plate 9. Ganesa from Jl. Mayor Ruslan, Palembang

The most recent, major, Image from Palembang, 1t is seated in the Indian fashion with a makuta reminiscent of the early Pallava. Probably 12th-13th century. 19

2. The Early Second Millennium

The Musi River Basin

Apart from the earlier sites already mentioned, there are some twelve known ancient sites in this region, situated on the Musi, the Air Lematang, and the Air Rawas, which probably date to the early second millennium or a little earlier (Map II). Most of these sites have been known since the beginning of this century. Other than Schnitger’s work in the 1930s and Bronson’s survey in 1973/74, however, very little systematic research on them has been undertaken. The following comments are based on written sources, supplemented by my own observations.

Air Musi

1. Sungsang Although the oldest archaeological evidence for habitation comprises a few sherds of fifteenth and sixteenth century date recovered during a field trip in 1979, Sungsang is important in that it gives a valuable insight into the nature of early Malay riverine and estuarine settlements in the Palembang region and the archipelago. The highest parts of Sungsang are some four to five meters above the low tide mark. The settlement itself is almost entirely built on piles, with houses linked by wooden walkways. The mud around the supporting piles of the living area is strewn with a thick layer of broken roofing tiles, broken earthenware, and imported ceramics of nineteenth and early twentieth century provenance and other domestic rubbish. There is good reason to believe that this accumulation of rubbish may conceal much earlier deposits of domestic debris, especially as it is now considered that the coastline here has remained unaltered for several hundred years.36 Sungsang is strategically situated to control both shipping passing through the Selat Bangka (the route used in ancient times) and access to the one navigable mouth of the Musi. It is unlikely to have been left unoccupied by anyone interested in controlling the maritime trade of the region.

2. Upang A few sherds of twelfth to fourteenth century Chinese ceramics were recovered here in 1979. There has been considerable erosion of the river bank over the past twenty-five years, however, due to the passage of large vessels up and down the Musi. On the opposite, right bank of the river is the site of the grave of Demang Lebar Daun, the Kubu chief whose daughter married Sri Tri Buana, ancestor of the founder of . Villagers report the existence of an old earthwork in the same vicinity. This information has still to be checked out.

3. Palembang One of the most important references to archaeological exploration at Palembang, and apparently overlooked by scholars until quite recently, is

36. P-Y. Manguin, "Sumatran Coastline in the Straits of Bangka: New Evidence for Its Permanence in Historical Times," SPAFA Digest 3, 2 (1982): 24-28, 49. See also O. W. Wolters, "A Note on Sungsang Village at the Estuary of the Musi River in Southeastern Sumatra: A Reconsideration of the Historical Geography of the Palembang Region," Indonesia 27 (1979): 33-50. 20 a brief reference in the Oudheidkundig Verslag of 1929 to a survey carried out by Ir. H. Maclaine Pont. Maclaine Pont spent some time in the area and concluded that there were three major centers of occupation here: the modern European and ’s city; a Mojopahit/Chinese ’’city” dating to the thirteenth and fourteenth century; and a third, earlier center based on the Putih Kuku swamp to the north and northwest of Bukit Seguntang where several important fragments of statuary and large bricks had come to light.37 This statement should now be extended to include the complex at Karanganyar, south of Bukit Seguntang (above). From a point of view of locational analysis (Map III), the scattered remains at Palembang fall into three main groups: those to the west and south of Bukit Seguntang (including the hill itself) (No. 1-5, and 16); those in the center of the modern town between the Sei. Sekanak and the Sei. Palembang (No. 6, 7, and 8); and those to the east of the city in the area known as Palembang Lamo (No. 9, 10, and 11). In addition to these are three scattered locations: the site of the Talang Tuo inscription (15), some distance to the west of Bukit Seguntang; the so-called Gunung Mahameru, a small hill on the left (Ulu) bank of the Musi (13); and the site of a former earthwork on the Ogan (14). Historical sources attest, however, that many of the former inhabitants of Palembang lived on rafts anchored in the river itself. Any debris from this type of dwelling may well have been swept away by the flow of the river or have sunk deep into the mud at the river’s edge. The only statue in the Palembang area possibly datable to the twelfth or thirteenth centuries is perhaps the recently discovered Ganesa image seated in the Indian style from Jalan Mayor Ruslan, a short distance to the north of Candi Angsoke (Plate 9). The habitation sites at Air Bersih and Gede ing Suro have yielded only post-1300 material and Bukit Seguntang and Talang Kikim either earlier, tenth century, material or that from the late thirteenth century. The Ganesa image is therefore of particular interest, especially as it is in virtually perfect condition and was discovered face down at a spot which had previously been leveled from a small hillock when houses had been built there some time ago. These circumstances suggest that the image might have been deliberately buried by the Sumatran Malays before they abandoned Palembang for Tumasik and, eventually, Melaka. Were there events which took place in the twelfth or thirteenth centuries which changed the nature of settlement in Palembang? For by the end of the fourteenth century it had a sizeable, and apparently independent, Chinese population. Although it had ceased to be a center of Malay political power, Palembang continued to be remembered as the source of Malay culture from which the founders of Melaka drew their legitimacy. The changes may have been the result of the Javanese raid of 1390s, or may have resulted from earlier dis­ turbances. What little archaeological evidence there is for this period, together with ethnographical observations in the lower Musi and Batang Hari regions tend to affirm Professor Wolters’ observations that there were certain distinctions between Palembang and Jambi where, as we shall see, developments seem to have been affected by Javanese rather than Malay influences.

4. Telukkijing (Negeri) This is the site of a candi associated in legend with the former realm of Tambun Tulang at the confluence of the Batang Harileko and the Musi. At

37. See OV (1929): 148-49; ”Het Verkenningswerk." 21

Tabah, slightly downstream was a site defended by a ditch at which remains of bricks and ironworking were to be found.38 39 As far as I can ascertain, no one has visited this site in recent years and no date is attributable to it.

5. Sareka38 On the left bank of the Musi between Sekayu and Muara Lakitan, Westenenk reported the existence of a candi, known formerly as Negeri Pangkalan Cina on the Danunjung riverlet.40 41 Like Telukkijing, nothing more seems to be known about it.

6. Bingin Jungun Situated on the left bank of the Musi between Muara Lakitan and Muara Kelingi, Bingin, or more correctly Bingin Jungun, is the site of at least two candi and the findspot of the four-armed Avalokitesvara image (discussed above, p. 13) and an unfinished Buddha (Plate 10). The Avalokitesvara is probably from the remains of a candi located within the confines of a settlement which stood upon a hill, defended by an earthen rampart and ditch immediately below the present day kampung and separated from it by a stream known as the Taligawe. The unfinished, and presumably late, Buddha image which possibly dates from about the twelfth to fourteenth centuries may have come from Tebing Candi on the opposite, right bank of the Musi where, apparently, there was formerly a second candi. The site is now a heavily overgrown rubber garden, but fragments of large bricks may still be found amongst the undergrowth. According to local informants, Bingin Jungun (from Malay: Jong, a sea-going vessel) was formerly the head of navigation on the Musi, due to the existence of a cataract there, and thus ships had to be hauled up past the obstruction. A channel is said to have been blown through the rocks in the colonial period, allowing larger vessels to pass on up to Muara Kelingi, which then became the head of navigation on the river. I have not been able to ascertain if this information is correct, 41 but it would appear that Bingin Jungun was certainly of some importance in the early second millennium. A few sherds of thirteenth to fourteenth century Chinese stoneware, beads, and iron slag have come to light here.

38. Westenenk, "Boekit Segoentang," p. 224. 39. The name "Sareka" possibly comes from the Javanese sarakah, potter; R. J. Wilkinson, A Malay-English Dictionary (Romanized) (: Macmillan, 1959), p. 1022. I am not aware of any pot making in this vicinity at the present time. 40. Westenenk, "Boekit Segoentang," p. 225. 41. I had the opportunity of visiting Bingin Jungun in early 1983, immediately after the worst flood in living memory, when the Musi was still in spate. I enquired of the village headman regarding antiquities there and was told about the earlier site just below the modern village, where the remains of many large bricks were to be found in the undergrowth of a rubber garden. I eventually produced a picture of the two images which had been removed from Bingin some sixty years earlier. The headman turned to his neighbors and remarked, not without pride, "Kito anak Buddho." From the extent of the remains at Bingin Jungun it would appear to have been formerly of considerable political and religious importance in the middle Musi area. 22

A late, probably 13- seated Image.

Plate 11. Buddha Image from Solok Sipin, Jambi

Datable to the 8th century, it is the earliest known image from the Batang Hari river basin. (Courtesy: Museum Nasionalv Jakarta) 23

7. Muara Kelingi A few large bricks and sherds of thirteenth to fourteenth century Chinese stoneware, including fragments of Lung ch'uan greenware, have been recovered from the left bank of the Musi, immediately opposite the confluence with the Air Kelingi.42 43

Air Rawas

8. Sungei Tingkip Situated on the left bank of the Sungei Tingkip, a small tributary which joins the Air Rawas opposite Bingin Teluk, this site came to light in 1980 when a farmer was clearing forest land to plant . He discovered a completely undamaged standing Buddha image in post-Gupta style with its hands held in vitarkamudraW lying in the undergrowth amongst the scattered remains of large bricks and fragments of Chinese stonewares dating to the twelfth to fourteenth centuries. The site has not yet been surveyed or excavated but it would appear to be of some importance, particularly as it lies in an area of rich gold-bearing gravels. The site is protected on its northern side by the Sungei Tingkip, there may be the remains of an earthen rampart around the rest of the area.

Air Lematang

9. Babat This is the site of a candi reported by Westenenk.44 4546*

10. Modong This also is the site of a candi reported by Westenenk.48

11. Tanah Abang Here were three brick foundations, now apparently almost completely destroyed, one of which was a Siva shrine where a fragmentary bust of the god was recovered. Schnitger remarks that all three sites on the Lematang appear to have Saivite associations,48 a fact which may indicate that they are later, rather than earlier, remains.

42. In 1979, R. Wisnu Wijaya reported the presence of quantities of ceramic sherds at Belani, on the right bank of the Air Kelingi, some distance upstream from its confluence with the Musi. O. W. Wolters, "Studying Srivijaya," JMBRAS 52, 2 (1979): 11, n. 45. 43. The image is currently kept at the office of the Department of Education and Culture (Dep. P & K) at Lubuk Linggau, Kabupaten Musi Rawas in South Sumatra. According to an informant, place names ending in "kip" such as "Tingkip" indicate an association with earlier settlements inhabited by local chiefs. 44. Westenenk, "Boekit Segoentang," p. 225. 45. Ibid. 46. F. M. Schnitger, Oudheidkundige Vondsten in Palembang (Leiden: Brill, 1936), Bijlage C, p. 6. to

Map IV. Batang Hari Basin 25

Air Komering

12. Japara Here is the site of a stone-built rectangular foundation thought to be the base of a Hindu shrine from which an image was removed in 1885.47

Air Lalang

13. Kelobak, near Bay at This is the site of a candi on the Air Berau, a branch of the Bayat which joins the Lalang.48 49 The Lalang flows into the Banyu Asin and is not strictly part of the Musi basin. The candi is said to be some 10 km walk upstream from the present day village of Bayat Hilir, which is accessible along a track some 6 km from Simpang Bayat on the new Palembang-Jambi road. The Lalang at Bayat Hilir is still tidal. Palembang is also accessible by water.

Other Sites

There are a number of sites, such as those at Tanjung Raja on the Ogan and Kayu Agung and Ulak Beko on the Komering, information regarding which is not available to me at time of writing. I have therefore been unable to comment on them. Kota Kapur, on Bangka has yielded, in addition to the Srivijayan inscription, a small fragmentary four-armed mitered Visnu image & similar to two other small images recovered near Karawang in . The Kota Kapur site which is surrounded by an earthwork, is no doubt of considerable importance in the context of seventh century Srivijayan expansion, but has yet to be examined by professional archaeologists. Indeed, archaeologically the island of Bangka, now greatly disturbed due to the activities of miners, has not been paid the attention it deserves. Traces of early tin mining at Koba and former gold working at Mawang and were recorded as long ago as 1824. In the previous year remains of a large sampan of a then-unknown type were discovered some two meters below the surface at Tubali.50

The Batang Hari Basin

There appear to be three distinct phases of development in this region, indicated by a seventh century Srivijayan inscription at Karang Berahi on the Batang Merangin; a later, possibly tenth to fourteenth century phase at

47. Ibid., p. 5; Archaeology of Hindoo Sumatraf p. 4. See also F. D. K. Bosch, "Verslag van een Reis door Sumatra," OV (1930): Bijlage C, 133-57; and E. P. Tombrink, "Hindoe Monumenten in de Bovenlanden van Palembang, als Bron van gescheidkundig Onderzoek," TBG 19 (1870): 1-45. 48. Westenenk, "Boekit Segoentang," p. 225. 49. Willem F. Stutterheim, "A Note on a Newly Found Fragment of a Four-Armed Figure from Kota Kapur (Bangka)," Indian Art and Letters 11, 2 (1937): 105-9. 50. Anon., "A Short Sketch of the Island of Bangka," Journal of the Indian Archipelago 5 (1851): 225-91. (An article translated from the Dutch from Tijdschrift voor NeGrland's IndiS [hereafter TNI] 2 (1846): 262, 268.) 26

Muara Jambi; and late thirteenth to fourteenth century developments around Sungei Langsat on the upper Batang Hari, an area associated with the toponym and the Malayo-Javanese prince , who became ruler of . There are two main groups of sites in this region (Map IV), those on the lower Batang Hari between Koto Kandis and Jambi and those on the upper Batang Hari around Sungei Langsat (above). There are also a number of other sites on the Batang Tembesi, Batang Merangin, Batang Tebo, and Batang Bungo, about which very little is known.

Lower Batang Hari

1. Koto Kandis The chance discovery of a bronze dipalaksmi in late Cola style in the river bank at Koto Kandis in 1982 was the first indication that there was an early settlement in this lower branch of the Batang Hari, an area once thought to have been a large gulf up until about the fifteenth century. A visit to Koto Kandis in June 1983 revealed a scatter of thirteenth to fourteenth century Chinese stonewares extending along the left shore of the Kuala Niur for over 1.5 km, which, together with bricks, beads, and tiny fragments of gold scrap, attest the existence of a flourishing settlement here in late Srivijayan times. More recently a pipisan or grindstone for medicinal spices, bearing a brief inscription in Old Javanese script has also been found here.

2. Muara Kumpeh Ilir Site of a Dutch factory in the early eighteenth century, Muara Kumpeh Ilir occupies a strategic position on the Batang Hari below Muara Jambi. The Batang Kumpeh provides an ideal, sheltered anchorage just off the mainstream of the Batang Hari and was once known as the main port for the Jambi region. Sherds of thirteenth to fourteenth century Chinese stonewares, local pottery, and other cultural remains were discovered here along with the timbers of an ancient ship in September 1982. The site would therefore appear to have been of importance during late Srivijayan times and may have served as a port for Muara Jambi. 52

3. Solok Sakean (on the Batang Kumpeh) The Batang Kumpeh, between Muara Kumpeh Hulu and Penarikan would appear to have been canalized to join the Batang Hari with the natural drainage system of the Batang Kumpeh. At what date this work was undertaken is not clear. At Solok Sakean on the left bank of the Batang Kumpeh, Chinese stonewares have been found dating from the tenth to fifteenth centuries and later. Local farmers have reported brick remains in their fields and occasional finds of jewelry.5152 53

51. Edwards McKinnon, "Brief Note," p. 39. 52. E. Edwards McKinnon, "New Data for Studying the Early Coastline in the Jambi Area," JMBRAS 57, 1 (1984): 56-66. Reprinted in SPAFA Digest 5, 2 (1984): 4-8, 11, 19. 53. E. Edwards McKinnon and B. S. Dermawan, "Further Ceramic Discoveries at Sumatran Sites," Southeast Asian Ceramic Society, Transactions No. 8 (, 1981), p. 5. M UARA JAMBI "1

Map V. Muara Jambi: Site Map, after Nazir

1. Candl KotomahUgal 5. Candl Gudang Garam 10. Candl Astano 2. Candl Kedaton 6. Candl Gumpung 11. Candl Teluk [or Henapo C1no): Desa 3. Kandang Kerbau or Alunalun (site of an 7. Candl Unggi Kerningking Dalam Iatana?» with a defensive ditch) 8. Telago Rajo (kolam or artificial lake) 4. Candl Gedong 9* Candl Kembar Batu X. Henapo or small candl

N.B.s There are said to be a number of other brick ruins on the south (left) bank of the Batang Har1 in addition to that partially excavated et Desa Kemingklng Da lam. The exact location of these remains has not yet been Identified. There may be other remains In the Dano Lamo area where a Image was found 1n the 1930s. 28

4. Muara Jambi Muara Jambi is the most extensive and probably the most important archaeo­ logical site in Sumatra (Map V). It covers an area of some 1,000 hectares, transected by a system of both natural and artificial waterways. With its nine (at least) brick-built candi and some twenty menapo (smaller, candi-like brick monuments) it is the first identifiable late first/early second millennium urban complex to be discovered in Sumatra. 54 During the course of clearing the candi prior to restoration and the subsequent excavations carried out by the Indonesian Research Center for Archaeology, there have been numerous finds of broken statuary and other artifacts. The most important of these are: an exquisite, but unfortunately headless, Prajanaparamita image (Plates 12 and 13), similar in style and execution to that of Singasari, datable to the early/mid thirteenth century; a fragment of a black stone Thai Buddha in early Ayudya style from about the mid to late fourteenth century; and a bronze gong with an inscription in , datable to the year 1231. In addition, a wide range of high quality Chinese ceramics have been found, dating from the tenth to the fourteenth centuries, as well as Thai Sawankhalok wares, and a range of local earthenware recovered both as surface finds and from controlled excavations. Muara Jambi had extensive political and trading connections with Java, China, , India, and, in all probability, Arabia. The presence of both Javanese and Thai imagery at this site underlines its cosmopolitan character. It is here that we begin to glimpse a vast, widespread, power center, with an extensive royal ceremonial and religious complex at Muara Jambi itself, an important riverine harbor (and commercial center?) at Muara Kumpeh, some 40 km below Muara Jambi, and smaller clusters of habitation, essentially linked by water, scattered around the whole area between Muara Kumpeh Ilir, Muara Jambi, Muara Kumpeh Hulu, and along the Batang Kumpeh and the Batang Hari up as far as Solok Sipin in the modern city of Jambi. Chau Ju-Kua, writing in 1225, would have been describing this scene when he stated that "the people live scattered about outside the city, or on the water on rafts of boards covered over with reeds." There are still vestiges of floating settlements at both Palembang and Jambi, in addition to those which line the river banks and have spread out over the dry land. One difference which I have noticed between dwellings in Palembang and Jambi is, however, that, although in both areas houses are often built on piles, it is mostly on the Musi at the present time that they are placed directly into the water. In Jambi, houses are generally built atop the high river banks. I wonder if this is due to Javanese, as distinct from Malay, influence or a purely empirical measure to avoid floods? Considering the general state of* things at Muara Jambi, the site would appear to have been sacked, probably about the year 1377. One local legend attributes the destruction of Jambi to Thai forces led by a prince of Jambi,54 55 but whether the attacking forces were a coalition of Thai and Javanese, or Javanese alone is not known. Future historical research may, perhaps, provide a clue.

54. M. Nazir, Arkeologi Klassik Daerah Jambi (Jambi: Departemen Pendidikan dan Kebudayaan, 1980). 55. The legend concerning Thai involvement in Jambi is given in A. H. W. de Kock, "Legenden van Djambi," TNI 8, 4 (1846): 35-56. The of a canal is also mentioned, together with the names of Jambi and Muara Kompeh. I am grateful to Dr. Barbara Andaya for bringing this to my attention. 29

Plate 12. Prajnaparamita from Muara Jambi

Similer to the Prajnaparanlta from S1ngasar1> this probably dates from the early/m1d 13th century. Its presence, together with other Jevenese Imagery 1n the Betang Hari region, underlines the strong Javanese connection with Jambi. (Courtesy: Indonesian National Archaeological Service)

Plate 13. Prajnaparamita (detail)

The rear of the seated Image, showing the Intricate relief work of carving 1n Imitation of a . 30

Here, then, was the third and perhaps greatest of the three "wijaya’s"56 5758 of San-fo-ch'is Ku Kang or Palembang, the Old Harbor; Kedah or Kataha, on the west coast of Malaya; and Javanized Ch'ien pei or Jambi. Jambi was described as rich in gold. In its hinterland lie three of the richest gold-bearing regions of Sumatra; the rich gravels of the Tembesi/Rawas area; the alluvial deposits along the Merangin; and those of the . It may have been this wealth that attracted Javanese attention to the region.

5. Solok Sipin Reputedly the findspot of an early Buddha image (1.72 m in height, Plate 11), here in 1936, Schnitger reported the existence of a heap of bricks. More recently, Nazir has identified the site of a candi high up on the left bank of a former channel of the Batang Hari.57

Upper Batang Hari

1. Betung Bedara Findspot of part of a Buddha image, reputed to have come originally from Sungei Aro, further downstream, sa

2. Sungei Langsat This complex of Javanized sites at Sungei Langsat, Rambahan, Roco, and Padang Lawas are best dealt with briefly as a single entity. They are associated with the thirteenth to fourteenth century polity of Dharmasraya. The area has yielded several important images and inscriptions. From Padang Roco came an Amoghapasa portrait image with an inscription dated to the year 1208 Saka (1286), which was rededicated in 1343 by Adityawarman who later became king of Minangkabau. From Sungei Langsat came the gigantic image of Adityavarman as Bhairava, set up before he withdrew finally to Buo and the gold-bearing region of the Minangkabau heartland. The sites at Sungei Langsat would appear to represent a separate (as distinct from Melayu-Jambi) stage of Malayo-Javanese power in Sumatra.59 60

3. Batang Tebo Tanah Priok Findspot of a bronze Buddha image (H. 25 cm).

56. See Wolters, "Studying Srivijaya," p. 23. Professor Wolters considers that the toponym San-fo-ch'i was an abbreviated device used by the Chinese for designating a tripartite polity of which Palembang and Melayu Jambi were two constituent parts. 57. Schnitger, Archaeology of Hindoo Sumatra, p. 7; Nazir, Arkeologi, p. 29. 58. Schnitger, Archaeology of Hindoo Sumatra, p. 10; Nazir, Arkeologi, p. 29. 59. See Schnitger, Archaeology of Hindoo Sumatra, p. 8, and also L. C. Weste- nenk, "De Hindoe-Javanen in Midden- en Zuid Sumatra," in Congress voor de Taal, Land, en Volkenkunde van Java: Handelingen (Weltevreden: Albrecht, 1921), pp. 1-39. 60. OV (1921): 77. 31

4. Batang Bungo Muara Bungo (Teluk Kuali: Air Hitam) Findspot of a badly damaged Buddha image which came, presumably, from the site of a candi.81

5. Batang Tabir Rantau Limakapas Findspot of four small gold-washed bronze images comprising a Padmapani (lotus missing), an Avalokitesvara, a standing Buddha image in the covered mode, and a seated Buddha in bhumisparsamudra, recovered in 1975.82 It is not clear if any brick remains exist in this area.

6. Batang Merangin Pamenang A few kilometers downstream from Karang Berahi, site of the late seventh century inscription, there is said to be an ancient earthwork. A number of Chinese ceramics, notably thirteenth to fourteenth century white or gray glazed folded-rim bowls have been found in the area. The area upstream of Pamenang is reputedly rich in alluvial gold.

7. Kerinci Lolo Lolo is but one of several ancient sites located on or around Danau Kerinci, a lake set in an isolated, but fertile valley in the Bukit . The lake is drained by the Batang Merangin, a tributary of the Batang Hari. Lolo itself has yielded a variety of artifacts, including stone implements, potsherds, carnelian beads and bronze remains,83 and a bronze urn or flask.84 An exquisite bronze Padmapani88 has also come from the Kerinci area. Kerinci would appear to have been of some importance in ancient times. Not only did it lie on a route which connected the Minangkabau area, through the headwaters of the Batang Merangin, eastwards into upper Jambi, but important sources of gold were to be found in neighboring Marga Tanah Rendah88and through the mountains at Pangkalan Jambu (or Jambi), ®7 circumstances which may have 63646566676162

61. Nazir, Arkeologi, p. 29. 62. Ibid., p. 31. 63. A. N. J. Th. a Th. van der Hoop, "A Prehistoric Site Near the Lake of Kerinchi (Sumatra)," in Proceedings of the Third Congress of Prehistorians of the Far East, ed. F. N. Chasen and M. W. F. Tweedie (Singapore: Government Printing Office, 1940), pp. 200-204. 64. F. D. K. Bosch, "Een Bronzen Vat van Kerintji," OV (1922): 60-65. 65. Schnitger, Oudheidkundige Vondsten, p. 8. The findspot of this image is apparently not recorded. 66. J. Tideman, Djambi (Amsterdam: Kolonial Instituut, 1938), pp. 225-26. 67. Thomas Barnes, "Account of a Journey from Moco-Moco to Pengkalan Jambi, through Korinci in 1818," in Malay Miscellany (Benkulen, 1822). Barnes describes 32

Plate 14. Bronze Padmapani from Rantau Limakapas, Batang Tabir

One of four emailf goldwashed bronzes from Rantau Limakapas. (Courtesy: Department P & Kf Jambl) 33 impinged upon the wealth of Kerinci itself. Archaeological evidence, including megalithic remains, suggest that the Kerinci sites were inhabited over a long period. Certain of the potsherds from Lolo, in particular a fine, white unglazed earthware mentioned by Van der Hoop, have since been found elsewhere in Sumatra and Java and also other parts of Southeast Asia, indicating a connection between Kerinci and the outside world during the late Srivijayan period.

8. Batang Tembesi Sarolangun An unfinished Ganesa image (1.73 m in height) was recovered here, presumably from a candi, the site of which is no longer known.* 6869 70

Selat Melaka: Northeast Jambi

1. Batang Tungkal Kuala Tungkal (Kabupaten Batang Hari) A bronze Aksobhya image in bhumisparsamudra (14.5 cm in height) was reportedly recovered from Kuala Tungkal in the 1930s.69

Other Sites

There are, no doubt, other sites which have either escaped my attention or are, as yet, undocumented and await discovery as and when an intensive survey of the Batang Hari and its tributaries is undertaken. Such a survey is long overdue but will, no doubt, be immensely rewarding as my own brief trips along stretches of the Musi and Batang Hari have already proved.

What then do these scattered and often fragmentary remains mean in terms of political and social development? The question of the development of urban centers in island Southeast Asia has been addressed by Bronson and Wisseman79 and latterly discussed by Wheatley, who concludes that the "apparently anomalous paucity of archeological evidence . . . is a less intractable problem than it at first appears . . . the proposition that ancient settlements were not greatly dissimilar to indigenous (that is excluding colonial) ones in recent times . . . [implies] that the remains of early settlements are not likely to be overly prominent in the present-day landscapes of or Indonesia."71 The various routes available for descending to both Jambi and Palembang, with the times taken for each. Tideman, DJambi, p. 226, mentions Pangkalan Jambu, which I take to be Barnes’ Pangkalan Jambi. 68. Schnitger, Archaeology of Hindoo Sumatra, p. 10. 69. Ibid., Plate XI. 70. Bronson and Wisseman, "Palembang as Srivijaya." 71. See Paul Wheatley, Nagara and Commandery: Origins of the Southeast Asian Urban Traditions (Chicago: University of Chicago Department of Geography, 1983), p. 248. See in particular his Chapter 6, "Cities of the Early Malaysian World." 34

experience of Kota Cina, Muara Kumpeh (Suak Kandis), Koto Kandis, and Palembang itself confirms that this supposition is indeed correct. Numerous ancient settlements do exist. It remains only for us to recognize the often all-too- flimsy evidence that they present. The problem is basically one of environment. Many such early lowland settle­ ments stood in or at the edge of swamps, creeks, and rivers, and the sites have since been buried beneath accretions of peat or alluvium. They would appear to have been, for the most part, straggling "open" settlements without permanent defenses72 or defended only on the landward side by no more than a fence or pagar of thorny bamboo and a ditch. Walled cities such as those which exist in or China are the exception. The nearest Sumatran equivalent to such sites is probably that of Muara Takus, on the Batang Mahat in Riau province. Muara Takus is surrounded by the remains of an earthen wall. This complex, which extends to some 125 ha, dates from about the eleventh to twelfth century.73 Modern equivalents to these riverine environments would suggest that they could support relatively sizeable populations. Perhaps one of the nearest equivalents to these late first millennium, early second millennium, sites on the Batang Hari is to be found at in Selatan, where a large proportion of the population still lives either on the river itself

72. Neumann’s observation regarding settlement on the Pane and Barumun rivers in northeast Sumatra in the latter part of the nineteenth century are relevant. No doubt a review of other contemporary Dutch sources would reveal other relevant descriptions. He states: "On the Bila plateau of Garoga, the kampungs are still substantially fortified, surrounded with pallisades several meters thick and, where necessary, with ditches and situated on hilltops, such as Hadataran and Bulu Payung, or on the edges of ravines such as Garoga and Parsosoran, which are natural fortresses. Where the Government is established, such fortified places have been abandoned, such as Botung, si Minabun, Batu Gondit, Hurung etc., or they are deprived of their protection, such as si Mandiangin, Dolok Sai, Tambiski, Mandalasena, Tapus, Gunung Tinggi, Lobu Tayas, Aek Gambir, Huruba etc. On the plains one still sees, here and there, complete bamboo pallisades such as at Purba si Nomba, Batang Baruhar, Portibi Jae, and Gunung Tua where, little by little, these traces of former insecurity are disappearing. "The kampungs in the lower reaches of the river are very different from these. They are more open. The houses lie more widely scattered, often without any regularity so that with few exceptions they can scarcely be considered as kampungs. Frequently there are extensive gardens between the houses, even whole ladangs, which unmistakably break up the connection in a kampung. We should properly speak, therefore, only of ladangs. The natives label these settlements with this name also. "Whereas in the upper reaches of the river, the site of the kampung is dependent upon the availability of drinking water, one finds that in the lower reaches the kampungs are situated exclusively on the large rivers which form also a means of communication. The most important kampung in the whole river basin is surely si Jawi jawi on the left bank of the Bila, near its confluence with the Pane. This kampung comprises partly a long, double row of houses along the river. Downstream live the Chinese who have established themselves to trade or engage in the fishing." J. B. Neumann, "Het Pane- en Bila-stroomgebeid of het Eilan Sumatra," Tijdschrift van het Koninklijk Nederlandsch Aardrijkskundig Genootschap 2, 2 (1885): 9-133; 66. 73. Schnitger, Archaeology of Hindoo Sumatra, p. 11. 35

Plate 15. Wooden Image from Ulu Musi

This small wooden Image (approx. 40 cm in height) was formerly in the Rumah Bari, Palera- bang. It is interesting as it portrays a type of image which does not usually survive in the archaeological context. Undated, it would appear to have been connected with some kind of fertility rite. The face, with its pointed beard, is reminiscent of the Jalong bronze from the Chema valley in which is thought to date from the 9th century.

Plate 16. Buddhapad from Rejang Lebong, Bengkulu

A reminder of Buddhistic influences in the Ulu Musi area, this comes from an old village site, Seguring near in Bengkulu province. 36

or at the water's edge. Other sites in Kalimantan Selatan (and perhaps elsewhere in ?) are similar to those in southeastern Sumatra and may provide useful comparative data. Finds of major statuary and bricks obviously indicate religious centers of some importance. It would appear that these arose in connection with the estab­ lishment of political power, at or near the seats of local chiefs, who presumably also controlled trade and whose settlements acted as nodes in the exchange network, funneling hinterland produce such as gold, ivory, damar, and rotan to the major centers downstream. At the same time, they presumably acted as distribution cen­ ters for imports: salt, cloth, ceramics of indubitably higher quality than anything produced locally; glass, in the form of beads, and possibly copper and iron. Most such settlements would be expected to appear at the confluences of tributaries with the main stream. This seems to be true in most cases, conforming to Bronson's dendritic model of hinterland/coastal trade.74 There are, however, exceptions, such as Bingin Jungun which appear to have arisen due to some other factor, in this case an impediment in the river itself, which created a natural break in the journey up- and downstream. From the southwest coast of Sumatra and throughout the Musi and Batang Hari basins, there emerges from the end of the seventh century the suggestion of a dendritic mandala, with its roots set well back into the hinterland among the rich sources of alluvial gold and valuable forest products. The populations in these areas, the ulunezen or "people of the ulu" as they are sometimes described in Dutch reports, might not have been particularly dense, but small numbers of collectors could have ensured a steady trickle of produce finding its way into the mainstream of riverine trade and thence down to the harbor states.75 Those on the periphery were not untouched by what went on downstream, for it is interesting to note that forms of Buddhist persisted in the mountain valleys of Kerinci (Ulu Jambi) and in Rejang Lebong (Bengkulu) until about the middle of the nineteenth century. These mountain retreats were certainly not easy to reach, but they were accessible, as discoveries of imported ceramic wares testify. The rivers acted as the major means of communication, and as a source of fish for food. The lower and middle reaches could have supported what were, in earlier times, sizeable populations. The discoveries of the past fifteen years have brought with them the reali­ zation that environmental factors are of the utmost importance in studying the archaeology of this and similar regions. There are no walled cities such as those in mainland Southeast Asia, but human adaptability allowed striking political developments to take place in an environment which might seem, to the outsider, unreliable and inhospitable.

74. Ben Bronson, "Exchange at the Upstream and Downstream Ends: Notes towards a Functional Modal of the Coastal State in Southeast Asia," in Economic Exchange and Social Interaction in Southeast Asia: Perspectives from Prehistory, History and Ethnography, ed. Karl L. Hutterer, Michigan Papers on South and Southeast Asia No. 13 (Ann Arbor: University of Michigan, 1977), pp. 39-52. 75. For a study of this type of exploitation in Malaya, see J. L. Dunn, Rain­ forest Collectors and Traders. A Study of Resource Utilization in Modern and Ancient Malaya, Monographs of the Malaysian Branch of the Royal Asiatic Society No. 5 (Kuala Lumpur: M.B.R.A.S., 1975). Such a study is still to be carried out in Sumatra but would, no doubt, be of considerable use in helping to identify resource exploitation and to reconstruct ancient trading patterns in aromatic resins particularly in northeastern Sumatra.